This post by John Quiggin raises a very interesting idea. For one thing, I think there's much to be said for this voting system if democracies are supposed to represent the desires of the people (and hence account for the intensity and complexity of the preferences of voters, not just their primary direction). But beyond that, if we were to expect that parties will do what is in their best interest and seek their own survival and growth, a plausible case can be made that this is very much in the party's interest. I presume it hasn't been proposed because of the potential uncertainty (from the perspective of the party and its campaigns) it would introduce, and because Labour fears such a change could lead to the eventual adoption of proportional representation. But given the current political divides in the UK - for now, this would seem to be a system that would benefit the incumbent party.
Posted by armand at April 15, 2005 01:00 PM | TrackBack | Posted to PoliticsFirst, my bias: I generally support single transferable voting -- I think that *in practice* it does tend to better "represent the desires of the people" in a way that "account[s] for the intensity and complexity of the preferences of voters."
But, and this is a big but, single transferable voting is eminently game-able. If everyone votes his or her honest preferences, it reflects those very well. But if I have good information about other people's honest preferences (say through polling), I (and others like me) can elevate, say, my second-best choice and vote for it first in order to ensure that my least-favorite choice is not elected. If there were a lot of this kind of gaming the system, single transferable voting might be a disaster, if representation of honest voter preferences is the goal.
But, despite the fact that "bullet voting" is a strategic best move for people who support just one candidate on a list when they can vote for more than one, it seems that it's rare that people (on the whole) think that strategically about voting, and actually use bullet voting.
But, the Internet allows for a lot more strategy in voting than there ever was before (cf. vote trading).
So I'm not sure how I come out on the single transferable vote as a mechanism to better capture fine-grained voter preferences. Do you know of any studies that clear things up?
Posted by: arbitransom at April 15, 2005 03:40 PM | PERMALINKWell I'd be very cautious about the possibility of vote trading taking off. I mean it can certainly happen, and I've heard that there are people trying to organize trades in the UK right now (between Labour and the LibDems) but ... well, I haven't seen evidence of it affecting elections. Yet.
But I certainly agree that bullet voting is entirely reasonable and possibly utility maximizing.
As to your last two sentences I'd suggest turning to a text that examines all sorts of systems across a variety of types of countries and districts. One I liked in grad school was Electoral Laws and Their Political Consequences edited by Bernard Grofman and Arend Lijphart. That deals with scores and scores of issues.
Posted by: Armand at April 16, 2005 02:05 PM | PERMALINK